GABA A receptorsmediatetheactionofmanyclinicallyimportantdrugsinteractingwithdifferentbindingsites.Forsomepotentialbindingsites, no interacting drugs have yet been identified. Here, we established a steric hindrance procedure for the identification of drugs acting at the extracellular ␣1ϩ␤3Ϫ interface, which is homologous to the benzodiazepine binding site at the ␣1ϩ␥2Ϫ interface. On screening of Ͼ100 benzodiazepine site ligands, the anxiolytic pyrazoloquinoline 2-p-methoxyphenylpyrazolo[4,3Ϫc]quinolin-3(5H)-one (CGS 9895) was able to enhance GABA-induced currents at ␣1␤3 receptors from rat. CGS 9895 acts as an antagonist at the benzodiazepine binding site at nanomolar concentrations, but enhances GABA-induced currents via a different site present at ␣1␤3␥2 and ␣1␤3 receptors. By mutating pocket-forming amino acid residues at the ␣1ϩ and the ␤3Ϫ side to cysteines, we demonstrated that covalent labeling of these cysteines by the methanethiosulfonate ethylamine reagent MTSEA-biotin was able to inhibit the effect of CGS 9895. The inhibition was not caused by a general inactivation of GABA A receptors, because the GABA-enhancing effect of ROD 188 or the steroid ␣-tetrahydrodeoxycorticosterone was not influenced by MTSEA-biotin. Other experiments indicated that the CGS 9895 effect was dependent on the ␣ and ␤ subunit types forming the interface. CGS 9895thusrepresentsthefirstprototypeofdrugsmediatingbenzodiazepine-likemodulatoryeffectsviathe␣ϩ␤ϪinterfaceofGABA A receptors. Sincesuchbindingsitesarepresentat␣␤,␣␤␥,and␣␤␦receptors,suchdrugswillhaveamuchbroaderactionthanbenzodiazepinesandmight become clinical important for the treatment of epilepsy.

Application of 4-(aminomethyl)cyclohexanecarboxylic acid (tranexamic acid; TAMCA) to the central nervous system (CNS) has been shown to result in hyperexcitability and convulsions. However, the mechanisms underlying this action are unknown. In the present study, we demonstrate that TAMCA binds to the ␥-aminobutyric acid (GABA) binding site of GABA A receptors in membranes from rat cerebral cortex and does not interfere with N-methyl-D-aspartate receptors. Patch-clamp studies using human embryonic kidney cells transiently transfected with recombinant GABA A receptors composed of ␣1␤2␥2 subunits showed that TAMCA did not activate these receptors but dose dependently blocked GABA-induced chloride ion flux with an IC 50 of 7.1 Ϯ 3.1 mM. Application of TAMCA to the lumbar spinal cord of rats resulted in dose-dependent hyperexcitability, which was completely blocked by coapplication of the GABA A receptor agonist muscimol. These results indicate that TAMCA may induce hyperexcitability by blocking GABA-driven inhibition of the CNS.

A successful unified pharmacophore/receptor model which has guided the synthesis of subtype selective compounds is reviewed in light of recent developments both in ligand synthesis and structural studies of the binding site itself. The evaluation of experimental data in combination with a comparative model of the alpha1beta2gamma2 GABA(A) receptor leads to an orientation of the pharmacophore model within the Bz BS. Results not only are important for the rational design of selective ligands, but also for the identification and evaluation of possible roles which specific residues may have within the benzodiazepine binding pocket.

SUMMARYConflict procedures can be used to study the receptor mechanisms underlying the anxiolytic effects of benzodiazepines and other GABA A receptor modulators. In the present study, we first determined the efficacy and binding affinity of the benzodiazepine diazepam and recently synthesized GABA A receptor modulators JY-XHe-053, XHe-II-053, HZ-166, SH-053-2'F-S-CH 3 and SH-053-2'F-R-CH 3 at GABA A receptors containing α1, α2, α3 and α5 subunits. Results from these studies suggest that each compound displayed lower efficacy at GABA A receptors containing α1 subunits and varying degrees of efficacy and affinity at GABA A receptors containing α2, α3 and α5 subunits. Next, we assessed their anxiolytic effects using a rhesus monkey conflict procedure in which behavior was maintained under a fixed-ratio schedule of food delivery in the absence (nonsuppressed responding) and presence (suppressed responding) of response-contingent electric shock. Relatively non-selective compounds, such as diazepam and JY-XHe-053 produced characteristic increases in rates of suppressed responding at low to intermediate doses and decreased the average rates of non-suppressed responding at higher doses. XHe-II-053 and HZ-166 also produced increases in suppressed responding at low to intermediate doses, but were ineffective at decreasing rates of non-suppressed responding, consistent with their relatively low efficacy at GABA A receptors containing α1 and α5 subunits. In contrast, SH-053-2'F-S-CH 3 and SH-053-2'F-R-CH 3 produced only partial increases in suppressed responding and were ineffective on non-suppressed responding, consistent with their profiles as partial agonists at GABA A receptors containing α2, α3 and α5 subunits. These behavioral effects suggest that the anxiolytic and rate-reducing effects of GABA A receptor positive modulators are dependent on their relative efficacy and affinity at different GABA A receptor subtypes.

Classical benzodiazepines (BZs) exert anxiolytic, sedative, hypnotic, muscle relaxant, anticonvulsive, and amnesic effects through potentiation of neurotransmission at GABA A receptors containing a 1 , a 2 , a 3 or a 5 subunits. Genetic studies suggest that modulation at the a 1 subunit contributes to much of the adverse effects of BZs, most notably sedation, ataxia, and amnesia. Hence, BZ site ligands functionally inactive at GABA A receptors containing the a 1 subunit are considered to be promising leads for novel, anxioselective anxiolytics devoid of sedative properties. In pursuing this approach, we used two-electrode voltage clamp experiments in Xenopus oocytes expressing recombinant GABA A receptor subtypes to investigate functional selectivity of three newly synthesized BZ site ligands and also compared their in vivo behavioral profiles. The compounds were functionally selective for a 2 -, a 3 -, and a 5 -containing subtypes of GABA A receptors (SH-053-S-CH3 and SH-053-S-CH3-2 0 F) or essentially selective for a 5 subtypes (SH-053-R-CH3). Possible influences on behavioral measures were tested in the elevated plus maze, spontaneous locomotor activity, and rotarod test, which are considered primarily predictive of the anxiolytic, sedative, and ataxic influence of BZs, respectively. The results confirmed the substantially diminished ataxic potential of BZ site agonists devoid of a 1 subunit-mediated effects, with preserved anti-anxiety effects at 30 mg/kg of SH-053-S-CH3 and SH-053-S-CH3-2 0 F. However, all three ligands, dosed at 30 mg/kg, decreased spontaneous locomotor activity, suggesting that sedation may be partly dependent on activity mediated by a 5 -containing GABA A receptors. Hence, it could be of importance to avoid substantial agonist activity at a 5 receptors by candidate anxioselective anxiolytics, if clinical sedation is to be avoided.